Abstract: There is equitable distribution thinking in authorities of other pages linked to PageRank. As pages are different, the method has some limitation. This paper takes advantage of Web link structure, proposes an unequal way to treat the different pages when distributing authorities and call it improved PageRank. Experimental results show that IPR improves relation degree by about 90%, comparing to PageRank.

Key words: Web structure mining, page ranking, improved PageRank(IPR)

摘要： PageRank对所链接的网页的“权威值”存在平均分配的思想，由于互联网的网页是千差万别的，因此这种方法存在一定的局限性。该文利用了Web链接结构，提出了一种权威值不均衡分配的方法(IPR)，通过与PageRank算法相比，IPR的排序结果比PageRank提高了近90%的相关度。

关键词: 网页结构挖掘, 网页排序, 改进的PageRank

CLC Number: 

• TP393